Air pollution and preterm birth: comparing trimester average and repeated threshold exposure metrics in a North Carolina birth cohort, 2003-2015.

Background: Exposures to ambient air pollutants during pregnancy are associated with preterm birth (PTB) in studies that evaluate average pollutant concentrations during potential "windows of susceptibility," referred to as critical windows.

Objective: We evaluated associations between ozone (O₃) or particulate matter <2.5 micrometers (PM_2.5) during pregnancy and PTB, using continuous and categorical metrics of average exposure and repeated exceedances of threshold concentrations.

Methods: We analyzed data from North Carolina (NC) birth certificates from the NC Birth Defects Monitoring Program, limited to live, singleton births without birth defects and delivered from 2003-2015. The EPA's Fused Air Quality Surface Using Downscaling (fCMAQ) model was used to assign exposures for trimester-average and daily threshold exceedances for O₃ (daily 8-h maximum) and PM_2.5 (24-h average) at different daily threshold concentrations (O₃: 55-70 ppb; PM_2.5: 20-35 µg/m³) during trimesters 1 and 2 of pregnancy. Descriptive statistics, including weighted kappa statistics, were evaluated. Modified Poisson regression models were used to estimate adjusted risk differences (RDs) per 10,000 births.

Results: Among the cohort of 1,368,370 births, 8.4% were PTBs. The number of daily exposure exceedances ranged from 0 to 81 days over different threshold concentrations during trimester 1 or 2. There was slight to moderate agreement in risk associated with trimester average and daily threshold exposures exceedances, with decreased agreement among higher threshold concentrations. RDs ranged from -31 to 29 for average O₃ exposure and from -2 to 89 for repeated O₃ threshold exceedances. RDs ranged from 23 to 120 for average PM_2.5 exposure and from -227 to 118 for repeated PM_2.5 exposure.

Significance: This study provides insights into the heterogeneity in risk of PTB when assessing different air pollution exposure metrics.

Impact: This study investigates the impact of O₃ and PM_2.5 exposures during pregnancy on the risk of PTB, using measures based on both average concentrations and repeated exceedances of threshold levels. The findings address uncertainties identified in recent NAAQS reviews and suggests that repeated exposure metrics may capture differences in exposure that are not apparent when evaluating average exposure metrics.